Delay/Distruption Tolerant Networking (DTN) Working Group

Meeting: IETF 122 - Bangkok
Session: Monday II (120 minutes)
Date/Time (UTC): 17 MARCH 2025 / 06:00 - 08:00

Area: INT
Area Director: Erik Kline
Chairs: Ed Birrane, Rick Taylor
Secretary: Adam Wiethuechter

Introduction, Note Well, Milestones (05 mins)

Speaker(s): Chairs
Document(s): N/A

Packed agenda, any bash? -- None
Erik, not planning to running again for AD, if interest please reach
out!

Set of items being worked on.
Mix of status of items here, and new things under considerations.

CoAP over BP (10 mins)

Speaker(s): Carles Gomez
Document(s): draft-gomez-core-coap-bp-03

-02 presented in 121, new CoAP option for message aggregation
Validation of this option, so presenting to DTN
New section on proxying, orthognal to aggregation
New CoAP payload length option
Proxy usally an intermediary

Ed Birrane: namespace translation, can you say more about this in terms
of syncronization in delay envirornments?

Not bad as proxy may know how to deal with such requests, and may know
the translated name.

BPSec can protect all fields of CoAP over BP, OSCORE for proxies
Open question: any reason for both OSCORE and BPSec concurrently?
Observer: conditional-attributes draft mentioned

Rick Taylor: OSCORE vs BPSec, not thought in great depth. They offer
slightly different options. BPSec BPA to BPA. It would be a deployment
scenario decision. Having both good.
Rick Taylor (Chair): asking for adoption?

May need to discuss due to two WG involved.

Erik Kline: asked the question, what does CORE wang? Changes to ARPA do
not go to IANA but to IESG chair.

DTN Reference Scenarios (10 mins)

Speaker(s): Camillo Malnati, Felix Flentge
Document(s): N/A

demonstration and validation of BP and extensions
focus to next 3-10 years
Earth, Lunar and Mars
Topology, data rates, latency, error model
Scenario files in CSV

Ed Birrane: Jorge Amodio (chat); Why not considering the Lunar Base as a
local relay for the rover?
Multiple considered. We wanted on far side as more interesting and
challenging for BP extension.
Will be Yellow Book in CCSDS, next spring

Rick Taylor: great to put in yellow book, data format standardized?
Idea is as an official format but still converging

Felix Flentge (chat): files will be published, probably SANA
Marius Feldmann (chat): Will the source code of the network simulation &
emulation environment be provided under a FOSS license?
Felix Flentge: we will probably be able to provide tooling under FOSS

BPv7 Custody Transfer and Compressed Status Reporting (15 mins)

Speaker(s): Felix Flentge
Document(s): N/A

Update on this topic.
'opportunistic' custody
Immediate needs, small tightly managed networks or single missions
Orange Book should be published in 2025
Ealy versions prototyped
Bundle Sequence IDs; two forms
Handing within Admin Element, assignment is use case specific
Custody Transfer using an Extension Block - requested
Counterpart is Compressed Custody Signal; adminstrative record
Any node can request custody, decision to take or refuse is mission or
network specific

Ed Birrane: CTEB deletion of side block. Signing of block, and sense of
approved custodians? Overlaps with security.
Yes, but did not look into all of these cases

Compressed Bundle Reporting (CREB extension block)
What can you do? e2e in sequence delivery, e2e accounting and
reliability

Rick Taylor: Really interesting work. Conflated 3 thigns into on
document. Signaling is really good it its own right. Would welcome as
IETF document. Sequencing is interesting and a wider discussion to
explore. Problem: RFC9171 defined a unique bundle ID already (src id +
timestamp), do not understand why a new ID concept.

Large sequence easier and more effecient.
Rick Taylor: so new number to make list smaller.

Biran Sipos: consideration for compressed reporting with range with
gaps.
Felix: no, much more complex and less likely scenario.
Ed Birrane: good work, if possible to find the public work to post to
list.

Interplanetary DNS (10 mins)

Speaker(s): Scott Johnson
Document(s): draft-johnson-dtn-interplanetary-dns-03

Predicated on thought that IP networks will appear on other wordly
bodies. Coherent naming structure in all locations with all location
having IP networks deployed.
Distinct DNS root on each world
TLDs are identical but content vary by locality
Truncated set of data for off-world domains on each area as stub zones
Logical dedicated TLDs
IPN RRType -- now time to review and modify
Naming of any BP nework resource
Definition needs updating, would be nice to support 3-element ID scheme;
not just 2-element.
This group may want to optomize this record.
Asking for WG adoption

Erik Kline: (slide 4 requested). Not going to work with DNSSEC. Not the
right approach IMO not to hack DNS. I can take this to list. We have
CDNs, think should have IP addresses.
How does not break DNSSEC, just segments
EK: if you don't have the whole zone, can not do DNSSEC.
If no query has to happen, do not understand why broken?
Ed Birrane: gfetting questions, take to list due to time here.
Rick Taylor: Get what you are trying to, it is vital we need to get this
build block right. Technical questions to WG. 50% yeah! 50% ughh

Status of DTNMA (10 mins)

Speaker(s): Jenny Cao
Document(s): draft-ietf-dtn-amm-03, draft-ietf-dtn-ari-04,
draft-ietf-dtn-adm-yang-03, draft-ietf-dtn-amp-01

draft-ietf-dtn-adm-yang-03 also to be presented in NETMOD WG (Wednesday
II, 19 MARCH 2025, 06:00 - 08:00).

ARI draft mentions in AMM draft are now all informative

Rick Taylor: circulating the YANG stuff around to address any issues
before IESG review.
Eric Vyncke: About the YANG-related drafts, may I suggest an early
review by YANGDOCTORS ?

Status of WG & Personal Drafts (10 mins)

Speaker(s): Brian Sipos
Document(s): draft-ietf-dtn-bpsec-cose-06,
draft-ietf-dtn-eid-pattern-01, draft-ietf-dtn-udpcl-00,
draft-ietf-dtn-bp-sand-00, draft-sipos-dtn-edge-zeroconf

BPSec COSE Context is stable and trial implementations interoperating
EID Patterns remove DTN scheme until later when better understood
UDPCLv2 -- what features are good for core draft?

Rick Taylor: demux draft, very happy to stop working and put it into
UDPCL. We can have that conversation.
Also unidirection draft, overlapping but details matter. Prefer to avoid
ambiguities
BP SAND; first attempt at what this mechanism looks like. open ended
like other adverstisement like in MANET.

Rick Taylor: BTPU, more like SAND, have that conversations of generics.

That is point of the layering. IPND combined together, you can implement
messaging on any CLA. Worth having feedback if that is unclear.

Open points:
Cose Context -- doing update and seeing what people think.
EID patterns -- now shortcut defined, rather use text that absence
Rick Taylor: why not use a "+" symbol?
UDPCL and SAND -- prototyped to make sure do thing
Zero config draft exists and wants to treated in implementation and has
been proof of concepted

Bundle Transfer Protocol - Unidirectional (10 mins)

Speaker(s): Rick Taylor
Document(s): draft-taylor-dtn-btpu-00

Two drafts now
Lots of link layer protocols working in terms of frames
Share some characteristics: frame based, unidirectional (very
important), unreliable, point to multipoint
Partial framing is a link layer property
Due commonality, can make a CL protocol. "you can be a CLA if you
implement this CLP on top"
Keeping it simple
Sliding window sync
FEC as a second companion draft
Sender and receiver have a pre-agreed window size. Window is for
transfer not segments. Can skip ahead, after long silences.
FEC -- RFC6363. If you link layer does it, use them.
Lots of documents in queue around BTPU...

RFC4838bis (10 mins)

Speaker(s): Ed Birrane
Document(s): RFC4838

4838 is 18 years old! (it can drive)
Old authors/editors interested in developing a BIS
Topic 1: DTN and IP
Topic 2: QoS
Topic 3: Fragmentation

Please be vocal on mailing list on these.

Key Establishment in the Space Environment (15 mins)

Speaker(s): Benjamin Dowling (on behalf of Britta Hale)
Document(s): https://arxiv.org/pdf/2503.06785

Terrestrail comms based on certain assumptions that are broken in
interplanetary comms.
Keys used for authentication or confidentiality using long term
pre-shared keys
How do we refresh keys in a secure way?
TLS Key Update does not provide randomness
CKA might be a better choice
RFC9420
Add key agreement messages to BPSec using MLS

Rick Taylor: Is the intent of the team to implement a standard MLS NP
service and define a protocol for the messages, I would welcome the
effort.
Makes sense for immediate next steps.
Rick Taylor: suggest WG read this paper. For own good and wellbeing.

BP QoS Extensions (10 mins)

Speaker(s): Teresa Algarra
Document(s): N/A

Proposing a QoS block for users (UQEB).
Network QoS block, not standardized due to being very admin level
specific

Rick Taylor: flow marking as IETF document

Open Mic (05 mins)

Speaker(s): N/A
Document(s): N/A